Chromophore Function and Interaction in Escherichia coli DNA

or by photoreduction. Either FADH2 or the pterin chromophore in dithionite-reduced native enzyme can function as a sensitizer in catalysis. Various en...
0 downloads 0 Views 1MB Size
Download PDF
Biochemistry 1990, 29, 552-56 1

552

Chromophore Function and Interaction in Escherichia coli DNA Photolyase: Reconstitution of the Apoenzyme with Pterin and/or Flavin Derivatives? Marilyn Schuman Jorns,* Baoyu Wang, Susan P. Jordan, and Latika P. Chanderkar Department of Biological Chemistry, Hahnemann University School of Medicine, Philadelphia, Pennsylvania I91 02 Received June 13, 1989; Revised Manuscript Received August 25, I989

Escherichia coli, contains a neutral flavin radical (FADH') plus a pterin chromophore (5,lO-methenyltetrahydropteroylpolyglutamate) and can be converted to its physiologically significant form by reduction of FADH' to fully reduced flavin (FADH2) with dithionite or by photoreduction. Either F A D H 2 or the pterin chromophore in dithionite-reduced native enzyme can function as a sensitizer in catalysis. Various enzyme forms (EFAD,,, EFADH', EFADH,, EPteFAD,,, EPteFADH', EPteFADH,, EPte) containing stoichiometric amounts of F A D in either of its three oxidation states and/or 5,lO-methenyltetrahydrofolate(Pte) have been prepared in reconstitution experiments. Studies with EFAD,, and EPte showed that these preparations retained the ability to bind the missing chromophore. The results suggest that there could be considerable flexibility in the biological assembly of holoenzyme since the order of binding of the enzyme's chromophores is apparently unimportant, the binding of F A D is unaffected by its redox state, and enzyme preparations containing only one chromophore are reasonably stable. The same catalytic properties are observed with dithionite-reduced native enzyme or EFADH2. These preparations do not exhibit a lag in catalytic assays whereas lags are observed with preparations containing FAD,, or FADH' in the presence or absence of pterin. Photochemical studies show that these lags can be attributed to enzyme activation under assay conditions in a reaction involving photoreduction of enzyme-bound FAD,, or F A D H ' to FADH,. EPte is catalytically inactive, but catalytic activity is restored upon reconstitution of EPte with FAD,,. The results show that pterin is not required for dimer repair when F A D H z acts as the sensitizer but that F A D H , is required when dimer repair is initiated by excitation of the pterin chromophore. T h e relative intensity of pterin fluorescence in EPte, EPteFADH', EPteFAD,,, or EPteFADH, has been used to estimate the efficiency of pterin singlet quenching by FADH' (93%), FAD,, (90%), or F A D H 2 (58%). Energy transfer from the excited pterin to flavin is energetically feasible and may account for the observed quenching of pterin fluorescence and also explain why photoreduction of FAD,, or FADH' is accelerated by the pterin chromophore. An irreversible photobleaching of the pterin chromophore is accelerated by F A D H z in a reaction that is accompanied by a transient oxidation of F A D H 2 to FADH'. Both pterin bleaching and F A D H 2 oxidation a r e inhibited by substrate. ABSTRACT: Native D N A photolyase, as isolated from

E x p o s u r e of DNA to ultraviolet light results in the formation of cyclobutane dimers between adjacent pyrimidine residues. The presence of pyrimidine dimers in DNA has been associated with mutagenesis, carcinogenesis, and cell death. Pyrimidine dimers can be monomerized in a visible light requiring reaction catalyzed by DNA photolyase. Escherichia coli DNA photolyasc contains pterin and flavin. The flavin chromophore in the isolated enzyme is present as a blue neutral FAD' radical (FADH'), but the physiologically significant form of the enzyme in E . coli contains fully reduced flavin (1,5-dihydro-FAD, FADH2) (Jorns et al., 1984; Sancar et al., 1987; Payne et al., 1987). The flavin radical in the isolated enzyme can be reduced with dithionite or by photoreduction but is resistant toward oxidation (Jorns et al., 1987a,b). Formation of an enzyme-substrate complex stabilizes FADH2 against air oxidation and also quenches the chromophore's fluorescence in a reaction that is fully reversible upon dimer repair (Jordan & Jorns, 1988). The pterin chromophore in native enzyme is present as 5 , IO-methenyltetrahydropterolypolyglutamate [5,10-CH+-H4Pte(Glu),](Wang et al., 1988; Johnson et al., 1988; Wang & Jorns, 1989). The polyglutamate moiety in the native chromophore is not essential for binding. The isolated enzyme is depleted with respect to the pterin chromophore but can bind additional 5,1O-CH+-H4Pte(Glu),,or 'This work was supported in part by Grant G M 31704 from the Kntional Institutes of Health.

* To whom correspondence should be addressed. 0006-2960/90/0429-552$02.50/0

5,lO-methenyltetrahydrofolate(5,10-CH+-H4folate) (Wang & Jorns, 1989). Reduction of the pterin chromophore with borohydride yields 5-methyltetrahydropterolypolyglutamate [5-CH3-H4Pte(Glu),] in a reaction that is accompanied by a complete loss of the chromophore's visible absorption and fluorescence (Jorns et al., 1987b; Wang & Jorns, 1989). A similar loss of pterin absorption is observed upon exposure of the enzyme to black light, but the product formed in this photobleaching reaction is not 5-CH3-H,Pte(Glu), (Jorns et al., 1987b: Hamm-Alvarez et al., 1989). Studies with the physiologically significant form of E . coli photolyase show that either the pterin chromophore or FADH2 can function as a sensitizer in catalysis (Jorns, 1987; Sancar et al., 1987). It has been proposed that the enzyme can repair dimers via a pterin-independent pathway since turnover is not significantly affected by reduction of the pterin chromophore with borohydride (Jorns et a]., 1987b). In contrast, preliminary studies have suggested that FADHl may be required when dimer repair is initiated by excitation of the pterin chromo-

'

Abbreviations: FAD, flavin adenine dinucleotide; FAD,, oxidized FAD; FADH', blue neutral FAD radical; FADH,, 1,5-dihydro-FAD; DTT, dithiothreitol; 5,1 O-CH+-H,Pte(Glu),, 5,l O-methenyltetrahydropteroylpolyglutamate; 5,l O-CH+-H,folate, 5,10-methenyltetrahydrofolate; Pte, 5,10-CH+-H4folate or 5,10-CH+-H4Pte(Glu),; 5-CH3H,Pte(Glu),, 5-methyltetrahydropteroylpolyglutamate; 5-CH3-H4folate, 5-methyltetrahydrofolate; SDS, sodium dodecyl sulfate; Tris, tris(hydroxymethy1)aminomethane; EDTA, ethylenediaminetetraacetic acid.

0 1990 American Chemical Society

Apophotolyase Reconstitution with Pterin or Flavin phore (Jorns et al., 1987a). Evidence consistent with these proposals has been obtained in reconstitution experiments which have permitted us to prepare enzyme containing FAD in either of its three oxidation states (FADH2, FADH', FAD,,) in the presence or absence of pterin and also flavin-free enzyme containing 5,l O-CH+-H,folate as the only chromophore. Comparison of the spectral properties observed for these various enzyme forms shows that energy transfer from the excited pterin singlet to FAD,,, FADH', or FADH2 is energetically feasible and might account for the observed quenching of pterin fluorescence by the flavin chromophore and also explain why photoreduction of FAD,, or FADH' is accelerated by the pterin chromophore. The possible catalytic significance of energy transfer is discussed. EXPERIMENTAL PROCEDURES Materials. Phenyl-Sepharose CL-4B was purchased from Pharmacia. Oligo(dT)18was obtained from P-L Biochemicals. 5-Formyltetrahydrofolate was purchased from Sigma. Coenzyme F420 from a Methanobacterium strain was a generous gift from Dr. Ralph S . Wolfe. TPTpT$T was a generous gift from Dr. James Alderfer. Enzyme Purification and Assay. E. coli photolyase was purified as previously described (Jorns et al., 1987a). Unless otherwise specified, all handling of the enzyme was done under yellow light. Enzyme assays were performed a t 21 "C with UV-irradiated oligo(dT),, as substrate, similar to those described by Jorns et al. ( I 985), except that the assay buffer (50 m M Tris-HCI, pH 7.2, containing I O m M NaC1, 1.6 m M DTT, and 1 m M EDTA) contained DTT. Substrate was prepared as previously described (Jorns et al., 1985). For some experiments, the flavin in the enzyme (FADH' or FAD,,) was convcrted to FADH2 with 5 m M dithionite and assayed immediately. Protein concentration, determined by the Bradford method (Bradford, 1976), was used in calculating specific activity values. Preparation of Apoenzyme. All steps were done at 0 or 5 "C, according to a procedure similar to that described by Van Berkel et al. ( 1 988), except that buffer conditions were varied to optimize the yield of apophotolyase. An aliquot (900 pL) of native enzyme (-IO4 M) in complete PRE buffer (50 m M Tris, pH 7.4, containing 50 m M NaCI, 1.0 m M EDTA, 10 mM DTT, and 50% glycerol) was diluted with 3.0 m L of buffer A (50 m M potassium phosphate, p H 7.0, containing 1.7 M ammonium sulfate, 0.5 m M EDTA, 10 m M DTT, and 20% glycerol). Solid ammonium sulfate was added to a final concentration of 1.7 M. The sample was applied to a 1.O-mL phenyl-Sepharose CL-4B column equilibrated with buffer A. The enzyme's chromophores were removed by washing the column with buffer B (buffer A adjusted to pH 3.5 and saturated with KBr). The column was brought back to neutral pH by washing with buffer A. The apoenzyme was then eluted with buffer C (100 m M potassium phosphate, p H 7.0, containing 0.5 mM EDTA, 10 m M DTT, and 50% ethylene glycol). Reconstitution Studies. An aliquot (5 mL) of apoenzyme (=7 X 10" M) was incubated a t 6 "C in buffer C containing = 1 0-4 M FAD and/or 5,1 0-CH+-H,folate. 5,l O-CH+-H4folate was prepared as described by Rabinowitz (1963). In some experiments, excess sodium dithionite (0.03 M ) was added to keep the flavin reduced during the incubation. After 20 h, the sample was dialyzed against buffer A and then applied to a phenyl-Sepharose CL-4B column (1.0 mL) equilibrated with buffer A. The column was washed with buffer A. The reconstituted enzyme was eluted with buffer C and then dialyzed against complete PRE buffer.

Biochemistry, Vol. 29, No. 2, 1990 553

In some experiments, enzyme was isolated after reconstitution with oxidized FAD and then incubated at 6 "C with a 5-fold excess of coenzyme F420 or a 10-fold excess of 5,lOCH+-H4folate in complete P R E buffer containing 25% glycerol. After an overnight incubation, the sample (1.2 mL) was diluted with 2 m L of complete P R E buffer without glycerol. Solid ammonium sulfate was added to a final concentration of 1.7 M. The enzyme was applied to a phenyl-Sepharose CL-4B column and reisolated as described above. In other experiments, enzyme was isolated after reconstitution with 5,10-CH+-H4folate and then incubated at 6 "C with a 10-fold excess of FAD,, in complete P R E buffer. After 20 h, the glycerol concentration was reduced to 20% by dilution with complete PRE buffer without glycerol, solid ammonium sulfate was added to 1.7 M, and the enzyme was reisolated as described above. Spectroscopy. Absorption and fluorescence measurements were made with a Perkin-Elmer Lambda 3 spectrophotometer and a Perkin-Elmer Lambda 5 spectrophotofluorometer, respectively. Fluorescence measurements with free 5,lOCH+-H,folate, which is unstable a t neutral pH, were made immediately after dilution of an acidic stock solution. Reduction of enzyme-bound flavin with dithionite and low-temperature fluorescence and phosphorescence measurements were conducted similar to those previously described (Jorns et al., 1987b). Unless otherwise specified, photoreactions were conducted under anaerobic conditions at 6 "C in 50 mM Tris, p H 7.4, containing 50 m M NaCI, 1 .O m M EDTA, 1.6 m M DTT, plus 8.3% glycerol using the same light source (two black lights, Sylvania F15T8/BLB, 15 W ) used for catalytic assays. In some experiments, the two black lights were replaced by two yellow lights (Westinghouse F15T8/GO, 15 W). The concentration of enzyme-bound FADH' was calculated on the basis of its absorbance at 580 nm (€580 = 4.8 X IO3 M-' cm-') (Wang & Jorns, 1989). The extinction coefficient of FADH' at 380 nm = 6.0 X IO3 M-I cm-l) was determined based on the absorption spectrum observed for reconstituted enzyme containing only FADH'. Extinction coefficients for enzyme-bound FAD,, (cM3 = 1 1.2 X lo3 M-I cm-I; €380 = 1 1.O X lo3 M-I cm-' ) were determined on the basis of the amount of free FAD (6450 = 11.3 X IO3 M-' cm-I) released upon denaturation of enzyme reconstituted with FAD,, alone (see Figure 3). The extinction coefficient of enzyme-bound pterin [5,10-CH+-H4Pte(Glu), or 5,10-CH+-H4folate](€380 = 25.9 X lo3 M-' cm-I) was determined in studies with native enzyme or enzyme reconstituted with 5,1 O-CH+-H4folateplus FAD,, after correcting the observed absorbance of the enzyme at 380 nm for the contribution due to FADH' or FAD,,, respectively. In these experiments, the pterin content of the enzyme was determined on the basis of the amount of free pterin [e360 = 25.1 X lo3 M-' cm-I for free 5,10-CH+-H4folate (Rabinowitz, 1963)] released upon denaturation of enzyme, according to the method described by Wang and Jorns (1989). To determine the extent of chromophore incorporation, protein concentration was determined on the basis of the absorbance of the reconstituted enzyme a t 280 nm, as described by Wang and Jorns (1989), except that in some experiments absorbance measurements were made with intact rather than denatured enzyme since AZs0was found to be unaffected by denaturation. (The ratio, A280,mtact enzymelA280,denatured enzyme varied from 0.99 to 1.03 in studies with native enzyme or enzyme reconstituted with FAD,, alone, FADH, alone, or FAD,, plus 5,lOCH+-H4folate.) RESULTS Effect of Thiols on Catalytic Assays with Native Enzyme.

554

Jorns et al.

Biochemistry, Vol. 29, No. 2 , 1990 n rn

u260

300

400

500

600

700

Wavelength (nm)

4

8 12 Time (min)

16

4

8 12 16 Time (min)

20

F I G U R E I : Catalytic activity with various photolyase preparations. The increase in absorbance at 260 nm, due to repair of dimers in UV-irradiated oligo(dT),8,is plotted versus time of irradiation with photoreactivatinglight. Unless otherwise noted, assays were conducted under standard conditions ( I .6 m M DTT). Data are normalized to the same protein concentration. Panel A: Lines 1-3 were obtained with native enzyme in the presence of 1.6 mM, 40 pM, or 5 NM DTT, respectively. Panel B: Line 1 was obtained with reconstituted enzyme containing only FADH'. Line 2 was obtained after reduction of FADH' to FADH, with dithionite. Panel C: Line 1 was obtained with cnzymc reconstituted with FAD,, plus 5,10-CH+-H4folate.Line 2 was obtained after conversion of FAD,, to FADH' by reduction of the reconstituted enzyme with dithionite followed by air oxidation. Panel D: Line 1 was obtained with enzyme reconstituted with FAD,, alone. Line 2 was obtained after reduction of FAD,, to FADH, with dithionitc.

FADH2 in dithionite-reduced native enzyme acts as a sensitizer i n catalysis whereas FADH' in the isolated enzyme is photochemically inactive with respect to dimer repair (Jorns, 1987; Sancar et a]., 1987). Nevertheless, dimer repair is readily detected in assays initiated with the isolated enzyme. Enzyme activity is typically measured a t saturating light intensity in the presence of saturating substrate [UV-irradiated oligo(dT),,] by following the increase in the absorbance at 260 nm that accompanies monomerization of thymine dimers (Jorns et a]., 1985). The facile photoreduction of FADH', observed with photolyase in the presence of thiols (Heelis & Sancar, 1986). suggested that a photochemical activation of the isolated enzyme might occur under standard assay conditions which include I .6 mM dithiothreitol (DTT). Under standard conditions a plot of dimer repair versus time is linear. The presence of a lag is difficult to detect even when initial data points are collected at I -min intervals. However, the existence of a small lag is suggested by the fact that plots of dimer repair versus time do not pass through the origin but rather intersect the x axis at about 0.4 min (Figure l a , line 1 ) . Although the rate of dimer repair was unaffected, plots that passed through the origin were obtained if the flavin radical was reduced with dithionite just prior to assay or if the DTT concentration in the assay buffer was increased to 17 m M (data not shown). At DTT concentrations lower than 1.6 mM, an initial lag was clearly discernible with untreated native enzyme and the rate observed after the lag decreased as the thiol concentration decreased (Figure l a , lines 2 and 3). Assays with di-

FIGURE 2: Absorption spectra of native enzyme and reconstituted enzyme containing FADH' plus 5,10-CH+-H4folate.Spectra of native (curve 1) and reconstituted (curve 2 ) enzyme were recorded at 5 "C in complete PRE buffer containing 25% glycerol. The inset shows the absorption spectrum of apoenzyme in 100 mM potassium phosphate, pH 7.0, containing 0.5 mM EDTA, 10 mM DTT, and 50% ethylene glycol.

thionite-reduced enzyme were unaffected by the DTT concentration in the assay buffer. The results suggest that native enzyme is activated under assay conditions by photoreduction of FADH' to FADH, in a DTT-dependent reaction. Apoenzyme Preparation. Native photolyase was bound to a small column of phenyl-Sepharose CL-4B at neutral pH in the presence of ammonium sulfate. Flavin and pterin were removed by washing the column with pH 3.5 buffer containing ammonium sulfate plus KBr. The apoprotein was recovered in 85-90% yield when the column was washed with neutral buffer containing 50% ethylene glycol. It exhibited a single absorption maximum at 280 nm (Figure 2, inset) and was catalytically inactive, but could be reactivated by adding back the removed chromophores (vide infra). The procedure used to prepare apoenzyme is similar to a method recently described by Van Berkel et al. (1988) for the reversible dissociation of flavoenzymes into apoprotein plus free flavin. We were unable to prepare reconstitutable apophotolyase using a variety of other published methods for the preparation of apoflavoproteins (e.g., acid ammonium sulfate precipitation, potassium bromide dialysis, treatment with guanidine hydrochloride, etc.). Apophotolyase could be prepared in 70% yield by passing a solution of the enzyme through a DEAE column at pH 3 in the presence of 50% glycerol. However, yields of reconstituted enzyme were poor ( ~ 5 % although ) the properties of the recovered material were similar to that observed when apoenzyme was prepared by the hydrophobic column chromatography method (Chanderkar and Jorns, unpublished observations). Reconstitution with FADH, plus 5,10-CH+-H$olate. The yield of the reconstituted enzyme was 67%, based on the amount of starting apoenzyme. Oxidation of FADH2 to FADH' occurred during isolation of the reconstituted enzyme, similar to that observed during purification of native enzyme (Sancar et al., 1987; Payne et al., 1987). The reconstituted enzyme (EPteFADH') contained nearly equimolar amounts of FADH' and 5,10-CH+-H4folate (Table I). The pterin content of the reconstituted enzyme was significantly greater than that observed for native enzyme which contains about 0.5 mol of pterin/mol of flavin (Table I). Native and re-

Biochemistry, Vol. 29, No. 2, 1990 555

Apophotolyase Reconstitution with Pterin or Flavin

Table I : Properties of Various Photolyase Preparations chromophore composition (mol of chromophore/mol of protein) enzyme preparation pterin" FADH' FADOX FADH, specific activity (%) lag in assay (min) native enzyme 0.54 0.92 100 EPteFADH') (see Figure 1 and Table I) suggested that photoreduction of FADH' might be faster than FAD,, and that both reactions might be accelerated by the pterin chromophore. Evidence to evaluate this hypothesis was sought by measuring the rate of flavin photoreduction under reaction conditions as close as possible to a standard assay. However, photoreactions were conducted under anaerobic conditions at 6 "C, substrate was

-0.8

I 4

8 Time fminl

12

Time (min)

FIGURE 8: Photoreactions with native enzyme. Reactions were conducted under anaerobic conditions at 6 "C in 50 mM Tris, pH 7.4, containing 50 mM NaCI, 1.0 mM EDTA, 1.6 mM DTT, plus 8.3% glycerol. Panel A: The sample was irradiated with black light. Flavin radical reduction was monitored at 580 nm. Pterin photobleaching was estimated from absorbance changes at 380 nm after correcting for the contribution due to flavin reduction. Lines 1 and 2 show first-order plots of absorbance changes at 580 nm and corrected absorbance changes at 380 nm, respectively. Panel B: The flavin radical was reduced with yellow light, and then the sample was irradiated with black light. Absorbance changes observed during black light irradiation at 380 and 580 nm are plotted in lines 1 and 2, respectively. Absorbance changes at 380 nm were corrected for the contribution due to FADH2 oxidation.

omitted, and the photoreaction buffer included a small amount of residual glycerol (8.3%) from the enzyme storage buffer which had no effect on assays. Anaerobic conditions were chosen, even though assays are conducted aerobically, because enzyme-bound FADH2 is stable against air oxidation in the presence of substrate (Jordan & Jorns, 1988). Photoreactions were measured at 6 "C owing to enzyme instability in the absence of substrate or substantial amounts of glycerol at the normal assay temperature (21 "C). A lag lasting about 1 min was detected when assays were conducted with native enzyme in standard assay buffer a t 6 OC, but surprisingly, the rate of dimer repair was unaffected by the decrease in temperature. Linear first-order plots were observed for photoreduction of EFADH' or EFAD,,. The reaction with EFADH' ( k = 0.22 min-I) was 2-fold faster than EFAD,, ( k = 0.1 1 min-I), consistent with the 2-fold difference in the length of the lag observed in catalytic assays (Table I). On the basis of initial rates, photoreduction of FADH' in native enzyme ( k = 1.1 min-I) was 5-fold faster than EFADH' whereas photoreduction of FAD,, in EPteFAD,, ( k = 0.22 min-I) was 2-fold faster than EFAD,,. The results are consistent with the proposal that flavin photoreduction is accelerated by the pterin chromophore. However, linear first-order plots were not observed for flavin photoreduction with native enzyme (Figure 8A, line 1) or with EPteFAD,, (data not shown). Both reactions were accompanied by a slower photobleaching of the pterin chromophore. Pterin photobleaching results in a large decrease in absorption at 380 nm ( A ~ 3 s o= 25.9 X lo3 M-' cm-' 1, as

Apophotolyase Reconstitution with Pterin or Flavin compared with smaller decreases due to FADH, formation from FADH' (Atjso = 1.1 X IO3 M-' cm-' or FADOX(AC38.0 = 6.1 X IO3 M-' cm-' ). Except for a n initial lag, linear first-order plots were obtained for pterin photobleaching with native enzyme ( k = 0.29 min-I) (Figure 8A, line 2). Qualitatively similar results were obtained with EPteFAD,,, except that pterin photobleaching was slower ( k = 0.077 min-') and the initial lag was more pronounced (data not shown). Although flavin photoreduction in native enzyme or EPteFAD,, is accelerated by the pterin chromophore, this rate acceleration is expected to diminish over the course of the reaction, owing to the accompanying photobleaching of the pterin chromophore. This suggested that linear first-order plots might be observed for flavin photoreduction in these preparations if the effect of the pterin chromophore could be eliminated. T o test this hypothesis, FADH' photoreduction in native enzyme was measured under yellow light (500-750 nm, A,, = 585), which excites only FADH', instead of the normal black light (300-450 nm, A,, = 355 nm), which excites both chromophores. In agreement with the predicted behavior, linear first-order kinetics were observed with yellow light ( k = 0.99 min-I). [No attempt was made to correct for differences in light intensity or spectral overlap of the light source with the absorption spectrum of the FADH'. However, the efficacy of yellow versus black light mediated FADH' reduction can be estimated by comparing the rate observed for the yellow light photoreduction of FADH' in native enzyme with that observed for black light photoreduction of EFADH' ( k = 0.22 min-I).] In a separate experiment, native enzyme was irradiated with yellow light and then exposed to black light to determine whether pterin photobleaching might be affected by prior reduction of FADH' to FADH,. Prior treatment with yellow light eliminated the initial lag in first-order plots but did not otherwise affect the rate of pterin photobleaching ( k = 0.26 min-I). However, pterin photobleaching was accompanied by an unexpected transient formation of FADH' which reached a maximum level after 3 min of irradiation, corresponding to a 17% reoxidation of FADH, formed in the yellow light reaction (Figure 8B). The results suggest that when native enzyme is irradiated with black light, without prior exposure to yellow light, the actual initial rate of FADH' photoreduction is probably faster than the observed rate, owing to the competing reoxidation of FADHz that accompanies pterin photobleaching. Pterin photobleaching is observed with flavin-free enzyme, but the reaction with EPte ( k = 0.050 min-I) is 5-fold slower than native enzyme. The results indicate that FADH, is not an obligatory electron donor but that it does accelerate the photobleaching of the pterin chromophore. The rate of dimer repair observed with EFAD,, was 80% of the rate observed when the flavin was reduced with dithionite just prior to assay (Table I), suggesting that 80% of EFAD,, must be photoreduced to EFADH, during the initial 12-1 4-min lag. This is consistent with photochemical studies where 75% of EFAD,, was reduced in 14 min. In the case of EFADH', the catalytic data suggested that 75% of EFADH' would be photoreduced during the initial 6-min lag (Table I), in good agreement with photochemical experiments where 73% reduction was observed in 6 min. In contrast, discrepancies between photochemical and catalytic studies were found for pterin-containing enzyme preparations. In the case of native enzyme, it was expected that photoreduction of FADH' would be complete in less than 1 min since almost no lag is detected under standard assay conditions and reaction rates are unaffected by reduction with dithionite just prior to

Biochemistry, Vol~29, No. 2, 1990 559 assay. However, photochemical studies with native enzyme showed that only 56% reduction of FADH' occurred in 1 min. Similarly, on the basis of catalytic data obtained with EPteFAD, (Table I), it was expected that 91% reduction of FAD, would occur in 6 min whereas only 60% reduction was observed in photochemical studies. The discrepancy between the catalytic and the photochemical data might be explained if photobleaching of the pterin chromophore was prevented by substrate since flavin photoreduction is slower in the absence of the pterin chromophore and is also decreased by the competing reoxidation of FADH, which accompanies pterin photobleaching. T o test this hypothesis, the flavin radical in native enzyme (1.8 X M) was photoreduced with yellow light, oligonucleotide substrate (1 .I m M TpTpTpT) was added, and then the mixture was irradiated with black light. Formation of FADH' was completely suppressed during the first 4 min of irradiation, and only a small amount of pterin photobleaching (6%) was detected. Substrate was apparently consumed during this period since further irradiation resulted in reaction rates similar to that observed in a control experiment without substrate. When the experiment was repeated with a 2-fold smaller amount of substrate, formation of FADH' was suppressed during the first 2 min of irradiation and pterin photobleaching exhibited a pronounced lag lasting about 1.5 min. The results show that pterin photobleaching and reoxidation of FADH, are inhibited by substrate. DISCUSSION The hydrophobic column chromatography method recently developed by Van Berkel et al. ( 1 988) was easily adapted to give good yields of reconstitutable apophotolyase. Enzyme containing equimolar amounts of pterin and flavin could be prepared by reconstituting the apoprotein with 5,10-CH+H,folate plus FAD,, or FADH,. Stoichiometric flavin or pterin incorporation was also observed upon reconstitution of the apoprotein with flavin (FAD,, or FADH,) in the absence of pterin or with pterin (5,10-CH+-H4folate) in the absence of flavin, respectively. Enzyme reconstituted with FAD,, or 5,10-CH+-H4folate alone retained the ability to bind the missing chromophore. The results show that the order of binding of the enzyme's chromophores is unimportant, that the binding of the flavin chromophore is unaffected by its redox state, and that it is possible to prepare reasonably stable preparations of reconstituted enzyme that contain only one chromophore. The results suggest that there could be considerable flexibility in the biological assembly of holoenzyme and that neither the pterin nor the flavin chromophore has a major structural and/or stabilizing function in native enzyme. The pterin chromophore in E. coli photolyase is replaced by coenzyme F,,, in photolyase from A. nidulans. Although these enzymes show a high degree of amino acid sequence homology (Yasui et al., 1988), an attempt to bind coenzyme F,,, to the pterin site in E . coli photolyase was unsuccessful. Pterin-Free Enzyme. Reconstitution experiments have permitted us to prepare pterin-free enzyme containing flavin in either of its three oxidation states. EFADH, is oxygensensitive but can be generated by reduction of EFADH' or EFAD,,. EFADH' is isolated after reconstitution with FADH, and can also be prepared by reduction of EFAD,, followed by air oxidation. The spectral properties observed for EFADH' or EFADH, are similar to that observed in studies with native enzyme when the visible absorption of the pterin chromophore was eliminated by borohydride reduction (Jorns et al., 1987b). Absorption spectra observed for EFAD,, or EFADH' exhibit features similar to those observed for the corresponding free flavins in nonpolar solvents, consistent with

560

Biochemistry, Vol. 29, No. 2, I990

a hydrophobic binding site for flavin in photolyase. EFADH,, prepared via reduction of EFADH' or EFAD,,, did not exhibit an initial lag in catalytic assays whereas pronounced lags were observed with EFADH' or EFAD,,. Photochemical studies showed that these lags could be attributed to enzyme activation under assay conditions in a reaction involving photoreduction of EFADH' or EFAD,, to EFADH,. That EFAD,, cannot repair dimers unless reduced to EFADH, is not due to an inability to bind substrate since formation of an EFAD,,-substrate complex was readily detected by the perturbation of the enzyme's visible absorption spectrum. The same turnover rate was observed with native enzyme or with EFADH, prepared via reduction of EFADH'. A 2-fold slower rate was observed when EFADH, was prepared via reduction of EFAD,,, suggesting that better recovery of enzyme activity is achieved via reconstitution with FADH, versus FAD,,. The results provide strong support for the proposal that enzyme containing only FADH, is fully competent catalytically. Formation of an enzyme-substrate complex with EFADH, quenched the fluorescence of the reduced flavin in a reaction that was fully reversible upon exposure of the complex to photoreactivating light, similar to that observed in studies with dithionite-reduced native enzyme (Jordan & Jorns, 1988). On the other hand, formation of an enzymesubstrate complex with EFAD,, did not affect FAD,, fluorescence. The results suggest that the observed quenching of FADH, fluorescence is not an "artifact" caused by substrate binding (e.g., a substrate-induced conformational change that affects flavin fluorescence) but is rather an indication that the singlet state of FADH, ('FADH2*) functions as an intermediate in catalysis. It has been proposed (Jordan & Jorns, 1988) that 'FADH,* donates an electron to the pyrimidine dimer, generating FADH' plus an unstable dimer radical anion (T$T'-) which rapidly monomerizes to yield T'- + T. The catalytic cycle is completed by electron return from T o - to FADH', regenerating FADH,. Flavin-Free Enzyme. The binding of 5,l 0-CH+-H,folate to apophotolyase causes several changes in the properties of the chromophore. Free S,IO-CH+-H,folate is unstable in aqueous solution at neutral p H owing to a ring opening reaction to yield IO-formyltetrahydrofolate (Kay et al., 1960: May et al., 1951). This hydrolytic reaction is suppressed in EPte, similar to that observed for native enzyme. A pronounced bathochromic shift of the absorption maximum of 5,10-CH+-H4folate(from 360 to 387 nm) is observed upon binding to apophotolyase, accompanied by a 67-fold increase in chromophore fluorescence. Although borohydride reduction converts either enzyme-bound or free S,IO-CH+-H,folate to 5-CH,-H4folate (Jorns et al., 1987b; Wang & Jorns, 1989), attempts to observe a comparable photobleaching reaction with free 5. IO-CH+-H,folate have not been successful (Wang and Jorns, unpublished observations). No dimer repair was detected with EPte whereas reconstitution of EPte with FAD,, yielded an enzyme preparation, EPteFAD,,, with catalytic properties similar to that observed when the apoprotein was reconstituted with pterin plus FAD,, in a single step. In each case, an initial lag was observed in assays with EPteFAD,, which could be eliminated by prior reduction with dithionite to generate EPteFADH,. The results shou that flavin is essential for catalytic activity and provide the first direct evidence in support of a proposal (Jorns et al., 1987a) that light absorbed by the pterin chromophore does not result i n dimer repair unless FADH, is also present.

Jorns et al.

I

I

I

Wavelength (nm)

Spectral overlap between flavin absorption versus pterin fluorescence as a function of the flavin redox state. Absorption spectra observed for enzyme-bound FAD,,, FADH', and FADH, are shown in curves 1-3, respectively. Curve 4 is the emission spectrum of the enzyme-bound pterin chromophore. F I G U R E 9:

Evidence for Pterin-Flauin Interaction. Photobleaching of the pterin chromophore is accelerated by FADH, in a reaction that is accompanied by a transient oxidation of FADH, to FADH'. The latter observation suggests that photobleaching may involve abstraction of an electron from FADH, by the excited pterin. FADH, is not an obligatory electron donor since pterin photobleaching can occur in flavin-free enzyme, although the rate is 5-fold slower. When EPteFADH, is irradiated in the presence of substrate, pterin bleaching and oxidation of FADH, are inhibited until the substrate is consumed. The effect of substrate may be related to the fact that oxidation of ground state FADH, to FADH' is suppressed upon formation of an enzyme-substrate complex (Jordan & Jorns. 1988) or may reflect formation of a common intermediate during catalysis and photobleaching. Photoreduction of FADH' or FAD,, is accelerated by the pterin chromophore (5- or 2-fold, respectively), as evidenved by comparison of initial rates of flavin photoreduction in the presence (EPteFADH', EPteFAD,,) or absence (EFADH', EFAD,,) of pterin. The fluorescence intensity of the pterin chromophore varies depending on the redox state of the flavin. The fluorescence intensity observed with flavin-free enzyme (EPte) was 14.4-fold greater than native enzyme (EPteFADH'), 9.6-fold greater than EPteFAD,,, and 2.4-fold greater than dithionite-reduced native enzyme (EPteFADH,). Since a similar shape was observed for the pterin fluorescence emission band in these preparations, the relative intensity of pterin fluorescence should be comparable to the relative fluorescence quantum yield. In this case, the data can be used to estimate the efficiency of pterin singlet (IPte*) quenching by FADH' (93%), FAD,, (go%), or FADH, (58%) (efficiency = 1 - fluorescenceEpt,n,,,,/fluorescenee,,,), assuming that the presence of flavin introduces a new path for 'Pte* deactivation but does not affect the rate of other deactivation reactions (e.g., emission, internal conversion). The observed quenching of pterin fluorescence by flavin and the accelerated rate of flavin photoreduction observed with pterin-containing enzyme might be due to energy transfer ('Pte* + flavin Pte + 'flavin*). Energy transfer, by either a dipole-dipole or an electron exchange mechanism, is unlikely unless the emission spectrum of the donor overlaps with the absorption spectrum of the acceptor (Turro, 1978). Figure 9 compares the fluorescence emission spectrum of EPte (curve 4) with absorption spectra observed for EFAD,, (curve l ) , EFADH' (curve 2), and EFADH, (curve 3). Spectral overlap is observed in each case,

-

Apophotolyase Reconstitution with Pterin or Flavin although the extent varies depending on the redox state of the flavin (EFADH' > EFAD,, > EFADH,). The results show that energy transfer from IPte* to flavin is energetically feasible. Energy transfer from 'Pte* to FADH, may be important during catalysis. Previous studies with EPteFADH, show that dimer repair can be initiated by excitation of the pterin chromophore (Jorns, 1987; Sancar et al., 1987) whereas our current studies show that FADH, is required for this reaction. Formation of an enzyme-substrate complex with EPteFADH, affects only the properties of FADH, (Jordan & Jorns, 1988), suggesting that FADH, is the chromophore that directly interacts with substrate. The observed quenching of FADH, fluorescence by substrate suggests that 'FADH2* is a catalytic intermediate. Energy transfer from 'Pte* would generate IFADH,* and provide a mechanism for utilization of light energy harvested by the pterin chromophore. The quantum yield determined for dimer repair with EPteFADH, (a = 1) (Jorns, 1987; Sancar et al., 1987) is higher than the estimated efficiency of IPte* quenching by FADH, (58%). However, quantum yield calculations were made using an extinction coefficient for FADH' in the isolated enzyme (e580 = 3.6 X IO3 M-' cm-') that has since been revised (€580 = 4.8 X lo3 M-I cm-l) (Wang & Jorns, 1989). A lower value for the quantum yield (a = 0.75) is estimated using the new extinction coefficient. The major path for dimer repair in EPteFADH, is likely to involve the pterin chromophore as sensitizer since the pterin chromophore absorbs most of the photoreactivating light. Pterin can be removed without affecting activity because activity measurements are made at saturating light and the light-harvesting step is not rate-determining under these conditions. ACKNOWLEDGMENTS We thank Dr. Ralph S . Wolfe for his generous gift of coenzyme FdZ0and Dr. James Alderfer for his generous gift of TPTpTfjT. Registry No. FADH2, 1910-41-4; FAD,,, 146-14-5; Pte, 1036012-0; DNA photolyase, 37290-70-3.

REFERENCES Bradford, M. (1976) Anal. Biochem. 72, 248-255. Eker, A. P. M. (1983) in Molecular Models of Photoresponsioeness (Montagnoli, G., & Erlanger, B. F., Eds.) pp

Biochemistry, Vol. 29, No. 2, 1990 561 109-132, Plenum, New York. Eker, A. P. M., Pol, A., van der Meyden, P., & Vogels, G. D. (1980) F E M S Microbiol. Lett. 8, 161-165. Hamm-Alvarez, S., Sancar, A,, & Rajagopalan, K. V. (1989) J. Biol. Chem. 264, 9649-9656. Harbury, H. A., LaNoue, K. F., Loach, P. A., & Amick, R. A. (1959) Proc. Natl. Acad. Sci. U.S.A. 45, 1708-1717. Heelis, P. F., & Sancar, A. (1986) Biochemistry 25, 8 163-8 166. Johnson, J. L., Hamm-Alvarez, S., Payne, G., Sancar, G. B., Rajagopalan, K. V., & Sancar, A. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 2046-2050. Jordan, S. P., & Jorns, M. S. (1988) Biochemistry 27, 8915-8923. Jorns, M. S. (1 987) in Flavins and Flavoproteins (Edmondson, D. E., & McCormick, D. B., Eds.) pp 233-247, de Gruyter, Berlin. Jorns, M. S.,Sancar, G. B., & Sancar, A. (1984) Biochemistry 23, 2673-2679. Jorns, M. S.,Sancar, G. B., & Sancar, A. (1 985) Biochemistry 24, 1856-1861. Jorns, M. S., Baldwin, E. T., Sancar, G. B., & Sancar, A. (1987a) J . Biol. Chem. 262, 486-491. Jorns, M. S.,Wang, B., & Jordan, S. P. (1987b) Biochemistry 26, 6810-6816. Kay, L. D., Osborn, M. J., Hatefl, Y., & Huennekens, F. M. (1960) J . Biol. Chem. 235, 195-201. May, M., Bardos, T. J., Barger, F. L., Lansford, M., Ravel, J. M., Sutherland, G. L., & Shive, W. (1951) J. A m . Chem. SOC.73, 3067-3075. Muller, F., Brustlein, M., Hemmerich, P., Massey, V., & Walker, W. H. (1972) Eur. J. Biochem. 25, 573-580. Payne, G., Heelis, P. F., Rohrs, B. R., & Sancar, A. (1987) Biochemistry 26, 7121-7126. Rabinowitz, J. C. (1 963) Methods Enzymol. 6, 8 14-8 15. Sancar, G. B., Jorns, M. S., Payne, G., Fluke, D. J., Rupert, C. G., & Sancar, A. (1987) J . Biol. Chem. 262,492-498. Turro, N. J. (1978) Modern Molecular Photochemistry, Benjamin/Cummings Publishing Co., Menlo Park, CA. Van Berkel, W. J. H., Van den Berg, W. A. M., & Muller, F. (1988) Eur. J. Biochem. 178, 197-207. Wang, B., & Jorns, M. S. (1989) Biochemistry 28, 1148-1 152. Yasui, A., Takao, M, Oikawa, A., Kiener, A., Walsh, C. T., & Eker, A. P. M. (1988) Nucleic Acids Res. 16,4447-4463.